Co-expression network analysis of biomarkers for adrenocortical carcinoma

Abstract

Adrenocortical carcinoma (ACC) is a rare malignancy with a poor prognosis. And currently, there are no specific diagnostic biomarkers for ACC. In our study, we aimed to screen biomarkers for disease diagnosis, progression and prognosis. We firstly used the microarray data from public database Gene Expression Omnibus database to construct a weighted gene co-expression network, and then to identify gene modules associated with clinical features of ACC. Though this algorithm, a significant module with R2 = 0.64 (P = 9 x 10-5) was identified. Co-expression network and protein-protein interaction network were performed for screen the candidate hub genes. Checked by The Cancer Genome Atlas (TCGA) database, another independent dataset GSE19750, and GEPIA database, using one-way ANOVA, Pearson's correlation, survival analysis, diagnostic capacity (ROC curve) and expression level revalidation, a total 12 real hub genes were identified. Gene ontology and KEGG pathway analysis of genes in the significant module revealed that the hub genes are significantly enriched in cell cycle regulation. Moreover, gene set enrichment analysis suggests that the samples with highly expressed hub genes are correlated with cell cycle. Taken together, our integrated analysis has identified 12 hub genes that are associated with the progression and prognosis of ACC; these hub genes might lead to poor outcomes by regulating the cell cycle. Copyright 2018 Yuan, Qian, Chen, Wu, Dan, Xiao and Wang.